/squarebot

SQuaRE Bot: the next-generation Kafka based Slack bot for Rubin Observatory.

Primary LanguagePythonMIT LicenseMIT

Squarebot

Squarebot is the nexus service for ChatOps and event-driven bots for Rubin Observatory. As a Slack app, Squarebot receives events from Slack and publishes them into specific Square Events Kafka topics in the Roundtable Kubernetes cluster. Other applications on Roundtable can consume these events and act on them, such as by posting messages back to Slack or by performing some other background automation.

flowchart LR
    subgraph Slack
    message[Slack message]
    interaction[Slack interaction]
    end
    subgraph sqrbot ["Squarebot"]
    eventapi["/slack/event"]
    interactionapi["/slack/interaction"]
    message -->|HTTP POST| eventapi
    interaction -->|HTTP POST| interactionapi
    end
    subgraph kafka ["Kafka Topics"]
    topicmention["lsst.square-events.squarebot.slack.app.mention"]
    topicchannel["lsst.square-events.squarebot.slack.message.channel"]
    topicgroup["lsst.square-events.squarebot.slack.message.group"]
    topicim["lsst.square-events.squarebot.slack.message.im"]
    topicmpim["lsst.square-events.squarebot.slack.message.mpim"]
    topicinteraction["lsst.square-events.squarebot.slack.interaction"]
    end
    eventapi --> topicmention
    eventapi --> topicchannel
    eventapi --> topicgroup
    eventapi --> topicim
    eventapi --> topicmpim
    interactionapi --> topicinteraction
    subgraph backends ["Backends"]
    backend1["Backend 1"]
    backend2["Backend 2"]
    end
    topicchannel --> backend1
    topicgroup --> backend1
    topicmention --> backend2
    topicinteraction --> backend2
Loading

Slack integration is implemented at the time. We plan to add support for other event sources, such as GitHub, in the future.

Squarebot is built on top of FastAPI, a modern Python web framework, with the Rubin/SQuaRE Safir library. Squarebot uses FastStream to publish messages to Kafka with the aiokafka library.

Development set up

Squarebot uses Python 3.12 or later.

You'll need nox to run the development environment:

pip install nox

Then you can set up a virtual environment with the development dependencies:

nox -s venv-init`

(If you already have a virtual environment, you can activate it and run nox -s init instead).

The tests require a local Kafka broker. If you have Docker, you can deploy a Kafka broker with Docker Compose:

docker-compose -f kafka-compose.yaml up

To run all linters, type checking, and tests, run:

nox

While developing and running tests, you can inspect messages produced to Kafka with Kafdrop at http://localhost:9000.

You can run individual sessions with nox -s <session-name>. See nox -l for a list of available sessions.

To tear down the Kafka broker, run:

docker-compose -f kafka-compose.yaml down